Anim8or Community

Anim8or disallows setting two object to face each other because it would create a circular dependency.

I think this should be changed somehow, because it creates a circular dependency only if they are in the same branch of a hierarchy tree.

In other words, if two objects aren't one parent of the other (whichever the level of "parent/child" you have), then those two objects can actually face each other, with no risk of circular dependency.

This isn't truly important, because this issue can be worked around by adding another object as a placeholder, but I wanted to point it out anyway, as it would be a minor improvement.

Uhm... this is sort of aetherial subject... more than speaking of avi materials, this is another workaround. I'll try to be as clear as possible, but the idea itself is cloudy, I hope I'll be able to explain it right.

So assume you want to add some kind of monitor to your a8r scene, and that you want it to display the frames of a movie. You would use avi materials if you could, otherwise you could do the trick using several materials and single frames from the movie.

In the attached zip there is the project and the pictures used to make the small gif animation attached separately.

The upper animation is made by sliding the "tape" in scene mode, the lower one is made with a morph target and a controller (by the way, once you set a controller for a morph target anim8or won't allow you editing it any more, at least it does so for me - clicking on its track doesn't pop up the edit key dialog).

So this approach has several negative sides.
- needs separate frames;
- needs a material for each frame;
- the tape that exceeds the viewport must be hidden by something else.

Variations of this technique can be made using an extruded regular polygon and sticking the frames to its faces, finally animating it by turning it about its center. Another variation could be to use single objects (simple planes) to store each image and then adding them all to a figure where, using one bone for each image, you can make each image to stand in front of all the others during each different frame.

The last approach is the most expensive as number of different things you need to set up in the project, but if you can automate a bit the process of manually editing a .an8 project file, this all becomes really fast. The most important advantage of using the last variation is that you don't need to hide the rest of the tape (the frames hide each other). In other words you can fit the animation in a flat plasma monitor inside of your scene

Here I come to the most interesting issue for me. The blue boxes sideways in the animation should hide the rest of the tape.

What if I could tell it that everything falling inside of that blue boxes must not be displayed/rendered?

It would need something like a material attribute tagging it as "empty space", "transparentizer" or whatever, and its result would be to ignore everything that falls inside of it. I think that this check could be done using the same "odd/even" test used when raster-filling an irregular poligon.

More or less in the same way, it would be a very useful animation feature to be able to set a material as an object exchanger.

Once again this is not so easy to explain, but I'll try.

Take the disc cutter I put in the WIP forum as example, and imagine the same work made in this way:

- object1: one object representing the white block.
- object2: one object (a figure or a morphed object) representing the hollow block and the disc.
- object3: one object, eventually a simple plane, set with a material tagged "changer" or whatever.

I put object1 in the scene and I put object3 between object1 and the camera.
I set a property of object3 telling it that object1 must become object2.

Now, when the renderer scans the lines of the viewport, it can tell if it is intersecting object3. If it does, it simply replaces object01 with object02 and renders that pixel as if the objects were actually exchanged.

I thought that this could lead to several problems. First of all, if object2 is not truly into the scene, it would share and use the tracks of object1. Fine until you don't want to use two figures and different animations.

I went to imagine this variation. Both object1 and object2 are put in the scene, and in the same spot of the scene. Object3 has two different special materials, one tagged A and the other tagged B. When the scanline intersect material A it displays only object1 and visaversa, crossing B only object2 will be visible. This is sort of adding a special kind of object to the whole system, as you'd need to tie object1 to material A *inside of object3 properties*.

I hope I've made it clear enough even for those who are not so accustomed to anim8or finesses and coding in general. I could make it all more transparent with some diagrams if needed.

Hi all,
this is a small animation I made while playing around with rigging. Also if some things I'd like to do aren't really easily feasible with the current rigging features of Anim8or (mechanical stuff such as the animated piston I mentioned elsewhere), nonetheless I've been able to make something nice, following the lesson about telescopic bones.

Here below you have a small xvid I put in a zip archive, and also two images, showing four frames from the video and the rigging of the mech arm.

More than a wip this is just a toy project, feel free to comment and criticize, I'm aware of a couple of weird things but I'd like to hear your opinion.

Hi all,
I've got a couple of suggestions about extending ASL features.

- Access / modify pivot's location and orientation;
I mentioned it elsewhere, the pivot is a very important reference point, way more evident than meshes' origins. By the way, it would be good to show meshes' origins in some viewmode (maybe just in pivot-edit mode while on wireframe view).

- Access to standard GUI functions;
(such as MergeFaces() and SubdivideFaces(), for instance, applied to selected items each time).

I'm asking for a specific reason: the polimagnet script used in "all directions" mode ends up enlarging faces. I'm going to code in that if a face gets too large, then it gets automatically subdivided, in order to add detail and volume. If I were allowed to call ConnectEdges(), tasks like this would become childplay.

Although I ask this for personal interest, I think that the access to those functions would be really useful for all ASL coders.

- Some kind of wait or delay function;
I mean giving back CPU ticks, not occupying them in a busy loop. The aim is to be able to put scripts in idle mode.

An example: the mirror script doesn't really need to duplicate the mesh all the times. Assuming also that Anim8or flags a mesh as "modified" after an user's action, the mirror script could handle it this way:

while($originalmesh.modified == false) {
    wait($idletime);
}
$originalmesh.modified = false; 

In a script like the polimagnet, checking if the meshes have been rotated or displaced would be enough for calling the wait() function. The "modified" flag is not truly necessary, but would be really useful anyway.

Have good time people.

Hi all,
here you can see a little render of mine.

Nothing really special I know, but since I like the final result I decided to share it.

This is how I did it:

1 - open an almost-all-quads face mesh
2 - select all edges
3 - connect edges
4 - switch to point-edit mode and invert selection
5 - delete points
6 - add edges where needed and fill holes
7 - select all edges and bevel them a bit
8 - extrude faces (with the extrude-connected tool)
9 - grow face selection
10 - apply different material to selected faces
11 - select all edges and bevel a bit more
12 - final result, ready to be subdivided

And here below I attach the pictures.

Feel free to use/modify this technique if you like it, you're welcome.

Hi all, here is a small attempt at doing a FOV zoom effect with Anim8or.

The effect is done by changing the distance and the FOV of the camera at the same time, the problem is that while the camera movement can be linear, the FOV change should be non-linear (logarithmic or geometric or whatever, I don't know).

The project and a small video are attached, just to get a quick idea if you don't have Anim8or under hands.

I guess that the FOV change should be made with a controller, but I'm already choking on ASL in object mode (see my post in the ASL section for that attempt). By now I tried to key it manually.

The aim is to change the perspective of the whole scene while keeping the size of the subject more or less constant (in this case the subject is the dodecahedron).

Take a look if you like.